We are excited! @initself.bsky.social & @rowansommers.bsky.social will present their **Umwelt Representation Hypothesis** at our next meeting. Sep 18, 11 am (CET) [Hybrid] Find the link to the online meeting here: www.cbs.mpg.de/en/cbs-coconut or join us at MPI CBS #neuroai #neuroskyence #ai

CBS CoCoNUT

CBS CoCoNUT: A new interest group that comes together every two weeks to have an informal exchange about cognitive computational neuroscience.

cbs.mpg.de

Victoria Bosch@initself.bsky.social · 3w ago

Are brains and artificial neural networks converging onto universal representations? There is a seductive idea making the rounds in NeuroAI / machine learning: train systems well enough, and they all converge on the same representation of reality (i.e. a unique world model). We have thoughts™ 1/n

📖 New review out today in @natrevneuro.nature.com: Towards an integrative neuroscience of metacognition 🧠🧪 Full-text link to the paper is here: rdcu.be/9YK3mF7dYhFa

This is our summary of where we think the field of Aphantasia research stands. Congenital aphantasia: what we (think we) know, what we thought we knew but were mistaken about, and what we don't know. www.sciencedirect.com/science/arti...

Congenital aphantasia: what we (think we) know, what we thought we knew but were mistaken about, and what we don't know

Congenital Aphantasia initially described a lifelong inability to visualise. But it is now understood to manifest as imaginative inabilities that can …

sciencedirect.com

What you see right now is not a livestream, but is delayed & edited by your brain 🧠. The brain uses a short time buffer to rewrite what you thought you saw using info that comes in later (postdiction). We study how (7T fMRI, retinotopy). Read our preprint: www.biorxiv.org/content/10.6... 👀 (1/N)

Super cool review on model identifiability in #NeuroAI! Generating “controversial” stimuli and using out-of-distribution (OOD) tests to reveal differences that standard benchmarks may hide. We’ve explored both directions in our recent work, two papers worth highlighting 🧵👇

Tal Golan@talgolanneuro.bsky.social · 3w ago

How can we design experiments that make computational models disagree? One section of our new @natrevneuro.nature.com Review with @kriegeskorte.bsky.social and @heikoschuett.bsky.social examines studies that used stimulus sets designed to elicit distinct predictions from competing models. 1/16

How can we design experiments that make computational models disagree? One section of our new @natrevneuro.nature.com Review with @kriegeskorte.bsky.social and @heikoschuett.bsky.social examines studies that used stimulus sets designed to elicit distinct predictions from competing models. 1/16

Are brains and artificial neural networks converging onto universal representations? There is a seductive idea making the rounds in NeuroAI / machine learning: train systems well enough, and they all converge on the same representation of reality (i.e. a unique world model). We have thoughts™ 1/n

The Umwelt Representation Hypothesis: rethinking Universality

Recent studies reveal striking representational alignment between artificial neural networks (ANNs) and biological brains, leading to proposals that all sufficiently capable systems converge on univer...

cell.com

Proud of this one, please check it out!

Victoria Bosch@initself.bsky.social · 3w ago

Are brains and artificial neural networks converging onto universal representations? There is a seductive idea making the rounds in NeuroAI / machine learning: train systems well enough, and they all converge on the same representation of reality (i.e. a unique world model). We have thoughts™ 1/n

New preprint www.biorxiv.org/content/10.6...: brain alignment in DNNs is almost fully complete at 3% of training. Features involved in classification are largely independent of those that predict the brain. With @niklasmuller.bsky.social , @irisgroen.bsky.social. @marcelvangerven.bsky.social

Good morning @ecvp.bsky.social! ☕If you are wondering how imagery and perception interact and compete, and how to quantify these competitive interactions in visual cortex with fMRI,🧠please drop by the poster session this afternoon and chat with me about my newest study (with @peelen.bsky.social)

Excited to share our preprint! w/ @martinhebart.bsky.social To understand how primates visually process objects in the world, we rely on both research in human and macaque IT. But what representations of object space are actually shared between them? biorxiv.org/content/10.6... Quick thread 🧵

Shared and Distinct High-Dimensional Object Spaces in Human and Macaque Inferotemporal Cortex

Human and macaque studies of inferotemporal cortex (IT) have shaped our understanding of object vision, yet the extent of their representational alignment and the precise nature of this correspondence...

biorxiv.org

The journal version is out! www.cell.com/trends-open/... AI is turning machine consciousness into a societal issue. Biological Naturalism (BN), the idea that only biological systems can be conscious, is gaining traction. We argue: either BN is untestable, or it is compatible with functionalism.

What biology can and cannot tell us about conscious AI

Progress in artificial intelligence is turning machine consciousness from a philosophical curiosity into a societal issue, and has led to criticism of the widespread computational functionalist framew...

cell.com

Adrien Doerig@adriendoerig.bsky.social · 4mo ago

1/n New preprint: What biology can, and cannot, tell us about conscious AI arxiv.org/abs/2606.02121 With @ulysse-klatzmann.bsky.social

Do you think in concepts? In pictures? A little bit of both? Thinking in pictures draws on parts of your brain also involved in seeing pictures: Early Visual Cortex (EVC). To learn how thinking in pictures may work, the amazing @noa tested 2 accounts in a neural network model. 🚨TWEEPRINT ALERT! 1/n

Man Taking Selfie with Flash Camera

ALT: Man Taking Selfie with Flash Camera

static.klipy.com

Excited to share that I'll be starting as a junior professor and Emmy Noether awardee at the University of Göttingen this October! 🥳 🚀 I'll be hiring #phd students and #postdoc across #machinelearning, #neuroscience and #cogsci: modirlab.github.io/open-positio... Please spread the word! 🙏

New position piece w/ @tylerbonnen.bsky.social out now in COBS! We suggest that data augmentation is a useful framework for understanding hippocampal contributions to generalization, and offers a path towards more precise modeling: 1/

I was interviewed for this youtube video on theories of consciousness (fair warning: it's in french) www.youtube.com/watch?v=TY25...

7+ théories de la conscience - Passe-science #70

YouTube video by Passe-Science

youtube.com

Excited to share our new paper: "Probing the content of semantic representations in body-selective regions" We propose a framework based on object co-occurrence to interpret semantic representations of natural scenes predicted by LLM embeddings. doi.org/10.1162/IMAG...

Probing the content of semantic representations in body-selective regions

Abstract. Recent advances in neural networks trained on natural language have revealed that category-selective regions encode complex semantics and contextual information of natural scenes in addition...

doi.org

Representation researchers: I think you definitely should check out this thought-provoking review by: @initself.bsky.social @rowansommers.bsky.social @adriendoerig.bsky.social @timkietzmann.bsky.social The Umwelt Representation Hypothesis: rethinking Universality pubmed.ncbi.nlm.nih.gov/42570893/

The Umwelt Representation Hypothesis: rethinking Universality - PubMed

Recent studies reveal striking representational alignment between artificial neural networks (ANNs) and biological brains, leading to proposals that all sufficiently capable systems converge on univer...

pubmed.ncbi.nlm.nih.gov

Online Now: The Umwelt Representation Hypothesis: rethinking Universality

The Umwelt Representation Hypothesis: rethinking Universality

Recent studies reveal striking representational alignment between artificial neural networks (ANNs) and biological brains, leading to proposals that all sufficiently capable systems converge on universal representations of reality. We argue that this claim of Universality is premature. We introduce the Umwelt Representation Hypothesis, which proposes that alignment arises not from convergence toward a single global optimum but from overlap in the ecological constraints under which systems develop. We review empirical evidence showing that representational differences between species, individuals, and ANNs are systematic and adaptive, which is difficult to reconcile with Universality. Finally, we reframe ANN model comparison as a method for mapping clusters of alignment in the ecological constraint space rather than as a search for a single optimal world model.

dlvr.it

(1/6) Do our visual neuroscience findings actually replicate? And do they generalize beyond the datasets they were found in? We've lauched re:vision, a community-driven initiative to answer these questions, and we are looking for scientistis to participate. re-vision-initiative.org

Replay of procedural memory is independent of the hippocampus 👀👀 www.nature.com/articles/s41...

Replay of procedural memory is independent of the hippocampus - Nature Neuroscience

Thompson, Rollik and colleagues show that, during sleep, the brain replays procedural memories in the striatum independently of the hippocampus. Replay content was shaped by prior reward and error out...

nature.com

Excited to share our review in @cp-neuron.bsky.social with @lauriebayet.bsky.social and @mickbonner.bsky.social! We describe how implementing principles from child development can advance the mechanistic plausibility and capacities of AI models We packed A LOT into this review, here's a quick 🧵

Neuron@cp-neuron.bsky.social · 2mo ago

Now officially published: our paper "Sensory horizons and the functions of conscious vision" w/ @smfleming.bsky.social, with the commentaries, and our response: Article: www.cambridge.org/core/journal... Commentaries: www.cambridge.org/core/journal... Response: www.cambridge.org/core/journal...

Now online (Neurocomputing), led by Nhut Truong, "Beyond topography: Topographic regularization improves robustness and reshapes representations in convolutional neural networks". Improvements found for image-noise, weight perturbations and adversarial attacks. www.sciencedirect.com/science/arti...

Excited to share that our paper has been accepted for a talk at #CogSci2026: Path Integration and Object-Location Binding Emerge in an Action-Conditioned Predictive Sequence Network Linda Ariel Ventura, Victoria Bosch, Tim C. Kietzmann, and Sushrut Thorat. Preprint: arxiv.org/abs/2602.03490. ⛓️

Path Integration and Object-Location Binding Emerge in an Action-Conditioned Predictive Sequence Network

Adaptive cognition requires structured internal models of objects and their relations. Predictive neural networks are often proposed to learn such world models, but how these are instantiated and how ...

arxiv.org

The hippocampus sits at the interface of perception & memory, allowing us to link past, present, & future through its role in generating predictions. Excited to share this theme issue in @royalsociety.org's Philosophical Transactions B, co-edited with @barense.bsky.social & @mariamaly.bsky.social

Volume 381 Issue 1954 | Philosophical Transactions of the Royal Society B | The Royal Society

Influential themed journal issues across the life sciences.

royalsocietypublishing.org

It was a delight to co-edit this issue with @peterkok.bsky.social & @barense.bsky.social, who are every bit as amazing to work with as you might expect 😊 I'm grateful to the authors, whose papers make this such an exciting collection, and to @catalinayang.bsky.social for the gorgeous cover!

Peter Kok@peterkok.bsky.social · 2mo ago

The hippocampus sits at the interface of perception & memory, allowing us to link past, present, & future through its role in generating predictions. Excited to share this theme issue in @royalsociety.org's Philosophical Transactions B, co-edited with @barense.bsky.social & @mariamaly.bsky.social

Separating the neural correlates of #consciousness from those related to unconscious processing is challenging. This #EEG study by @maelanmenetrey.bsky.social &co suggests two-stages of visual processing as individual stimuli are integrated into a conscious percept. 🧪 #AcademicSky plos.io/4wwzxNb

Both brains and RNNs can re-use components of computation across similar tasks or contexts. But what exactly are those “shared components”? How can they be used to solve several tasks? We address these questions in a new preprint with @avm.bsky.social! Link: www.biorxiv.org/content/10.6...

Interpretable compositional computation with recurrent neural networks

Flexible cognition utilizes reusable components to enable rapid adaptation of behavior to different contexts or tasks. Analysis of artificial neural networks trained on multiple tasks suggested that t...

biorxiv.org

Lovely to see the full range of excellent commentaries on our BBS article with @matthiasmichel.bsky.social, together with our response, now out: Target article: www.cambridge.org/core/journal... Commentaries: www.cambridge.org/core/journal... Our response: www.cambridge.org/core/journal...

Matthias Michel@matthiasmichel.bsky.social · last yr.

Very happy to announce that our paper “Sensory Horizons and the Functions of Conscious Vision” is now out as a target article in BBS!! @smfleming.bsky.social and I present a new theory of the evolution and functions of visual consciousness. Article here: doi.org/10.1017/S014.... A (long) thread 🧵